search

gokulkarthik / text2speech

Towards Building Text-To-Speech Systems for the Next Billion Users - Microsoft Research Intern Work - Accepted at ICASSP 2023
https://arxiv.org/abs/2211.09536
46 stars 8 forks source link

Punjabi #9

Open sultanovazamat opened 1 year ago

sultanovazamat commented 1 year ago

Hi! Thanks for the great job! I am wondering if Punjabi is also open-sourced, since it is not mentioned in your language list, but there are weights available for that language.

Thanks!

D1neshRawat commented 1 month ago

Hi Sultan, I am facing multiple problems while using the tts from gokulkarthuk.

How Did you used punjabi model ? I am trying to use it in this way. from TTS.TTS.api import TTS import torch import soundfile print(TTS().list_models()) BASE_DIR = os.path.dirname(os.path.dirname(os.path.abspath(file)))

Initialize TTS model with custom path model_path = "/Users/dineshrawat/pork/models/v1/as/fastpitch/best_model.pth" config_path = "/Users/dineshrawat/pork/models/v1/as/fastpitch/config.json" vocoder_path = "/Users/dineshrawat/pork/models/v1/as/hifigan/best_model.pth" vocoder_config_path = "/Users/dineshrawat/pork/models/v1/as/hifigan/config.json" tts = TTS(model_path=model_path, config_path=config_path, vocoder_path=vocoder_path, vocoder_config_path=vocoder_config_path).to(device_id) g Getting two kinds of error

when using the TTS from coqui-ai it gives speaker id is not correct. when using gokulkarthik tts lib i am not able to import the library and give error that. ModuleNotFoundError("No module named 'TTS.TTS.api'")